Debating Deposits: An Inter-Observer Variability Study of Lymph Nodes and Pericolonic Tumor Deposits in Colorectal Adenocarcinoma
Jonathan B Rock, M Kay Washington, Volkan Adsay, Joel Greenson, Elizabeth Montgomery, Marie Robert, Rhonda Yantiss, Amy Lehman, Wendy L Frankel. Ohio State University, Columbus, OH; Vanderbilt University, Nashville, TN; Emory University, Atlanta, GA; University of Michigan, Ann Arbor, MI; Johns Hopkins University, Baltimore, MD; Yale University, New Haven, CT; Weill Cornell Medical College, New York, NY
Background: The AJCC 7th Ed defines pericolonic tumor deposits (TD) as discrete foci of tumor in pericolic fat showing no evidence of residual lymph node (LN). This definition relies on subjective features rather than size (5th Ed) or shape (6th Ed) and includes the new category N1c. For staging data to be meaningful, pathologists must agree on features used to distinguish positive LN from TD. We evaluated interobserver variability and developed a list of features helpful in this distinction.
Design: Tumor metastases (25) from right-sided colorectal adenocarcinomas were selected where the distinction between a positive LN and a TD was challenging. Virtual slides were reviewed by 7 pathologists with an interest in GI pathology. A list of features thought helpful in distinguishing LN metastases was compiled and ranked for usefulness by all. Each metastasis was diagnosed as a positive LN or TD. For each case called a positive LN, features used in the distinction were listed.
Results: A “complete agreement” was found in 11 of 25 metastases (44%), of which 5 were deemed a positive LN and 6 a TD with a kappa statistic of 0.4 and 95% CI of (0.28, 0.67). In the remaining 14 cases, concurrence rates were 6/7 (4 cases), 5/7 (6 cases) and 4/7 (“non-agreement” 4 cases).
The top ranked discriminating feature was round shape followed by peripheral lymphocyte rim, peripheral lymphoid follicles, a subcapsular sinus, LN in surrounding fat and thick capsule. In practice, round shape was also the most utilized (89%) when diagnosing the 25 metastases followed by thick capsule, peripheral lymphoid follicles, peripheral lymphocyte rim and size >3 mm (each >60%).
Conclusions: Statistically significant agreement between evaluators was found, but some inconsistency remains. When the distinction between a positive LN and TD is difficult, we suggest using round shape, lymphoid follicles, peripheral lymphocyte rim, LN in surrounding fat, subcapsular sinus and thick capsule to aid in making a diagnosis.
Monday, March 4, 2013 9:15 AM
Proffered Papers: Section D, Monday Morning